E L S E V I E R Gene 205 (1997) 161-171

GENE A N I N T E R N A T I O N A L J O U R N A L O N

GENES AND OENOME~

Nuclear DNA diversity in worldwide distributed human populations

Ewa Zi~tkiewicz a, Vania Yotova a, Michal Jarnik .,1, Maria Korab-Laskowska a,2, Kenneth K. Kidd b, David Modiano c, Rosaria Scozzari d, Mark Stoneking e, Sarah Tishkoff b,3,

Mark Batzer f, Damian Labuda a,. a Centre de Recherche de rH6pital Sainte-Justine, Centre de Canc~rologie Charles Bruneau,

Dkpartement de Pbdiatrie de l' Universitk de Montreal, Montrbal, (Quebec), Canada H3 T 1 C5 b Yale University School o f Medicine, Department of Genetics, 333 Cedar St., New Haven, CT06510, USA

Fondazione Pasteur Cenci-Bolognetti, Istituto di Parassitologia, Universita 'La Sapienza', P. le A. Moro 5, 00185 Rome, Italy a Dip. Genetica e Biologia Molecolare, Universita "La Sapienza', P. le A. Moro 5, 00185, Rome, Italy

° Department o f Anthropology, Pennsylvania State University, University Park, PA 16802, USA f Department of Pathology, Stanley S. Scott Cancer Center, Louisiana State University Medical Center, 1901 Perdido Street,

New Orleans, LA 70112, USA

Accepted 14 July 1997

Abstract

Nucleotide variation was examined in an 8 kb intronic DNA bordering exon 44 of the human dystrophin gene on Xp21. Thirty- six polymorphisms (substitutions, small insertions/deletions and one (T)n microsatellite) were found using SSCP/heteroduplex analysis of DNA samples from mixed Europeans, Papua New Guineans as well as from six African, three Asian and two Amerindian populations. In this way the European bias in the nuclear polymorphism ascertainment has been avoided. In a maximum likelihood tree constructed from the frequency data, Africans clustered separately from the non-African populations. Fifteen polymorphisms were shared among most of the populations compared, whereas 13 sites were found to be endemic to Africans and four to non-Africans. The common sites contributed most to the average heterozygosity (H, =0.101% _+0.023), whereas the endemic ones, being rare, had little effect on this estimate. The FsT values were lower for Africans (0.072) than for non-Africans (0.158), suggesting a higher level of gene exchange within Africa, corroborating the observation of a greater number of segregating sites on this continent than elsewhere. The data suggest a recent common origin of the African and non-African populations, where a greater geographical isolation of the latter resulted in a smaller number of newly acquired polymorphisms. © 1997 Elsevier Science B.V.

Keywords: DNA polymorphism; Dystrophin; Human evolution; Ascertainment bias

I. Introduction

Disclosure of the content and organization of the genetic information in human D N A is a goal of the ongoing genome project. The linear sequence of nucleo-

* Corresponding author. Tel: +1 514 345 4931, ext. 3586/sec. 3282; Fax: + 1 514 345 4731; e-mail: Labuda@ere.Umontreal.ca 1 Present address: NIAMS, LSBR, Bethesda, USA 2 Present address: D6partement de Biochimie, Universit6 de Montr6al, Montr6al, Canada 3present address: Department of Biology, Pennsylvania State University, University Park, PA 16802, USA

Abbreviations: PCR, polymerase chain reaction; RFLP, restriction fragment length polymorphism; SSCP, single-strand conformational polymorphism.

0378-1119/97/$17.00 © 1997 Elsevier Science B.V. All rights reserved. PH S0378-1119(97) 00408-3

tides of the D N A message will soon be completely known. The second dimension of the human genetic heritage, its sequence variability within populations, has only started to be explored, both in qualitative and quantitative terms. According to the neutral theory (Kimura, 1983) most of the intra-specific variability at the molecular level is selectively neutral. It is maintained within the species by the balance between the mutational input and random extinction. The D N A polymorphisms represent a transient situation where one allele chosen at random is on its way to fixation, In the absence of selection the amount of genetic variation among subpo- pulations is shaped by mutation rate, random fluctua- tions in allele frequences (drift) and gene flow (exchange of alleles between populations), and strongly depends on demography (effective population size, migration

162 E. Ziftkiewicz et al. /Gene 205 (1997) 161-171

etc.). For this reason the genomic variability observed in the present-day human population is not only a record of the history of our genes but also of our species; knowing it better is crucial to solving the questions concerning the evolution of modern humans.

The knowledge of human DNA diversity originates to a large extent from studies using markers developed for linkage mapping and ascertained in a small number of samples, primarily of European origin. As a result of the European ascertainment, the human genomic vari- ability has suffered from the frequency bias of the polymorphisms tested (e.g. Bowcock et al., 1991; Mountain and Cavalli-Sforza, 1994; Rogers and Jorde, 1996). Furthermore, the use of RFLP markers was a source of sequence-based bias, since recognition sites of frequently used restriction enzymes often included CpGs of which the human genome is depleted (Bird, 1987; Labuda and Striker, 1989), and had an average G + C content that did not match the base composition of human DNA (see e.g. Cooper and Schmidtke, 1984; Cooper et al., 1985).

As a result of the limited sampling of individuals and/or populations as well as the differences in detection methods, the estimates of genome variability differed significantly. The density of segregating sites on the autosomes ranged from 1/370 to 1/108 (e.g. Murray et al., 1984; Cooper et al., 1985; Chakravarti et al., 1986; Hofker et al., 1986; Antonarakis et al., 1988; Fullerton et al., 1994), whereas the values reported for the X chromosome were from 1/732 to 1/190 (Cooper et al., 1985; Hofker et al., 1986; Zietkiewicz et al., 1995). The estimates of nucleotide diversity H, ranged from 0.1 to 0.39% on the autosomes (Murray et al., 1984; Cooper et al., 1985; Chakravarti et al., 1986; Hofker et al., 1986; Antonarakis et al., 1988; Li and Sadler, 1991; Fullerton et al., 1994) and from 0.04 to 0.09% on the X (Cooper et al., 1985; Hofker et al., 1986; Zietkiewicz et al., 1995).

To obtain unbiased information on the human genome diversity we investigated sequence variability in 13 globally distributed human populations. An 8 kb DNA segment of the human X chromosome was exam- ined using SSCP combined with the heteroduplex analy- sis, a direct and objective technique comparable to DNA sequencing but more cost-effective (Orita et al., 1989; Zietkiewicz et al., 1992). The analysis of the distribution of the new (non-ancestral) allele frequencies at the 35 segregating sites in all the populations studied provided us with new information, shedding light on the history of modern humans.

2. Materials and methods

2.1. Genomic samples

Human DNA samples from the existing collections in the authors' laboratories are listed in Table 1.

2.2. Dys44 sequence polymorphisms

We analysed a total of 7622 nucleotides out of the 8035 nucleotide-long dys44 sequence, spanning exon 44 of the dystrophin gene (cDNA positions 6499-6646) and its flanking introns between positions -2853 to - 1 upstream and 1 to 5034 downstream. The numbering refers to the GenBank sequence U94396, where the following modifications to the previously deposited sequence (M81257) were introduced: a G-to-A change at position 710 of intron 44, a one-nucleotide insertion at position 1754 of intron 44, and a 195-nucleotide 'insertion' including positions - 2347 to - 2541 (replac- ing a 10-nucleotide segment -2347 to -2356) in intron 43. To search for gel mobility variants the dys44 segment was divided into 30 fragments ranging from 92 to 480 bp, obtained by PCR combined where indicated with restriction enzyme digestion (Fig. 1 and Table 2). The amplification was in 20 lal, in 20 mM Tris HC1 (pH8.4) (at 25°C), 50mM KCI, 1.5mM MgCI2, 0.2 mM dNTPs (each), 0.2 lal (2 gCi) of ~-[32p]dCTP, with 1 jaM primers (each), 1 unit Taq DNA polymerase (BRL) and 5 ng DNA. Following denaturation (3 min, 94°C), the amplification was performed for 30-35 cycles (94°C for 30 s, annealing temperture as in Table 2 for 45 s and 72°C for 90 s), terminated at 72°C (10 min). The PCR products were examined using SSCP/ heteroduplex analysis, as described (Zietkiewicz et al., 1992). For typing length variants the PCR step at 72°C was 30 s rather then 90 s, the primer was radioactively labelled at its 5'-end (see Zietkiewicz et al., 1992 and below), and the electrophoresis was carried out under denaturing conditions. DNA mobility variants were characterized by direct dideoxy-sequencing of the PCR products (ABI 373A DNA sequencer).

2.3. Allele specific oligonucleotide, ASO, hybridization

Pentadecanucleotides were synthesized complemen- tary to each allelic variant identified, to serve as ASO- probes for DNA typing by dot-blot hybridization (Table 3). 2-8 kb (kilobase) dys44 fragments (depending on DNA sample quality) were amplified as above using combinations of primers from Table 2, except that the volume was 50 ~tl, the radioactivity was omitted, and a two-step PCR-program was applied (first 14 cycles were as above but the 72°C incubation was 120 s; in subse- quent 25 cycles the 72°C incubation time increased in 7 s increments per cycle). Each reaction was brought up to 100 gl with water, incubated at 94°C for 2 min and cooled on ice; the solution was then made 10 x SSC ( 1.5 M NaC1, 150 mM sodium citrate, pH 7) by addition of 100 gl 20 x SSC. Aliquots of 100 gl were then applied in parallel onto a Hybond@-N ÷ membrane (Amers- ham) to create two identical twin 48-dot blots. After rinsing with 100 gl of 10x SSC the membranes were

E. Zictkiewicz et al. / Gene 205 (1997) 161-171 163

Table 1 Population parameters for the 13 groups studied

Population Number of chromosomes S H,× 103 (+S.E.) H,~.,o × 103 H,,

African Americans 67 31 1.06 ( + 0.23 ) 0.91 0.813 Mossi (Burkina Faso) 25 24 1.05 (+0.23) 0.81 0.883 Rimaibe (Burkina Faso) 23 24 1.06 (_ 0.23) 0.79 0.786 Biaka Pygmies (Central Africa Republic) 89 21 0.98 (+0.23) 0.89 0.618 Nigerians 15 21 0.92 (+0.22) 0.76 0.558 M'Buti Pygmies (Congo) 58 18 0.82 (+ 0.21 ) 0.63 0.517 Europeans (mixed) 175 20 0.74 (+0.18) 0.74 0.394 Siberians a 24 19 0.93 (+0.22) 0.93 0.360 Papua New Guineans b 69 16 0.71 (+ 0.20) 0.70 0.205 Chinese 85 18 0.76 (+0.20) 0.76 0.131 Japanese 67 16 0.70 (+ 0.20) 0.69 0.061 Maya 80 16 0.84 (+0.22) 0.84 0.353 Karitiana (Brazil) 83 15 0.60 (_ 0.16) 0.60 0.281 World 860 35 1.01 (+0.23) 0.90 0.579 Africans 277 33 1.05 (+ 0.23) 0.84 0.773 Non-Africans 583 22 0.90 (+ 0.23) 0.90 0.270

S, Number of segregating sites (including substitutions and small indels); H,, average heterozygosity per nucleotide site (nucleotide diversity); H,Eu,, "H,' taking into account only the polymorphisms found in Europeans; H,., average heterozygosity of the (T), microsatelite; aTundra and Forest Nentsi; bCoastal and Highland.

immersed in 1.5 M NaC1, 0.5 M NaOH for 10 rain and in 1.5 M NaCI, 0.5 M Tris-HC1 (pH 7.2) for 15 min DNA was fixed by exposing the membranes to 254 nm UV at the energy of 120 000 ~ /cm 2 in Stratalinker 1800 (Stratagene).

Blots were pre-hybridized for 30 min (rotary oven) in 20ml l xSSPE (150mM NaC1, 10mM NaH2PO, 1.1 mM EDTA, pH7.4), 0.75M NaC1, 70mM Tris-HC1 (pH 7.4), containing 1% SDS and 200 ~tg/ml heparin, at hybridization temperature, TH (see Table 3). ASO probes, 50pmol, were 5'-labelled using ~-[32p]ATP (6000 Ci/mmol) and T4 kinase (Gibco BRL) to a specific activity of 1-3 × 106 cpm/pmol (250 000-750 000 cpm/ng). Hybridization with the 0.8-2.0 pmol ASO probe (~2 000 000 cpm) was carried

out for 40 min at TH. The membranes were then washed with 2 x SSPE, 0.1% SDS for 10min at room temper- ature, two times for 10 min at TH, rinsed with 2 x SSC at room temperature, and exposed overnight at -80°C with a screen. Identical twin membranes were probed by the allelic ASOs and always read in parallel. This compensated, if necessary, for varying concentrations of the individual amplified DNA samples and, by the same token, for the variance in the probe activity. In addition, DNA samples of known allelic content (SSCP variants of known sequence) served as positive controls for the allelic probes. After stripping (5 min in boiling 0.5% SDS) the membranes were stored at -20°C and reused (up to 12 times) for hybridization with other probes.

Exon 44

500 nt

6499,, / 6 6 4 6 • /

-2853 '", / +5034 I I I I i i , " ~ ( ' i I I I I I I I I I l l

T ?, T T _ _ T 1A 2 T T 4 A _ _ 5 ~ T A 8 Y 10 Y 12 __.J3B ,, 15

IB 3 4B 6 71] 9 11 13A 14

I I I I I I I J I If I I I , . i ~ I I I II i III I I I I I I I , , : J , ,

Fig. 1. Dys44 DNA segment. Flanking nucleotides -2853 and 5034 delimit dys44 region within the upstream and the downstream intron of dystrophin exon 44 (cDNA positions 6499 6646), respectively. Short horizontal lines below correspond to the overlapping PCR fragments (Table 2) used in the SSCP analysis, some of which were restriction digested (double-head arrows). The approximate distribution of the polymorphic sites (asterisks) is shown at the bottom.

164

Table 2 PCR amplification of dys44 fragments

E. Zif tkiewicz et al. / Gene 205 (1997) 161-171

Dys44 ~agment PCR primers

IA F AGG ~ATT TGT TGA AT R GCA CAG AGA AGT ACC AGT T

IB F CAAGAG TGG AGA CTG ATG G

R GAC AGC ATT TTG GTA GCA T

2 F CTC TTC GGC TAC CTT CG

R TGG GCA GAG AAA GGA ATT A

3 F GTG CCT GGC TAT TAG TAA

R CAT TTG GTC ACC TTC CAG T

4A F AAC ATA CAG CCC TGG TcC

R GTC ACA CTG TAC CCC ATA

4B F ACG AAT TAT TGA TTT ATT GG

R GTT CAT CAA CTG AAAGGA GT

5 F TGG TAA CTT TGT TCA TAT TA R GCG TAT ATT TTT TGG TTA TA

6 F CTT GAT CCA TAT GCT TTT ACC TGC A

R TCC ATC ACC CTT CAG AAC CTG ATC T

7A F TGA TTT GTT TTT TCG AAA T

R ACC TTG CTG TTA CGA TGC T

7B F AGC ATC GTA ACA GCAAGG

R CCG TGTAATAAA CAC AGT G

8 F TTG CTAAAT TAC ATA GTT TAG GC

R AGC CCAAAA TTC ACT TC

9 F GTA TGG ATT CCC AAT CTG

R TTA CAG AGA AAA GAG AGA CC

10 F TAG GCT CCT TAAAGT GCC R CTT ACA CTG TTC TGG TCA TAA C

II F TTT TAT GCT TTG TTA TGA CC

IIA F GCA TAC ATG AAC TGA TCA AG

R ATG AGG ACC ATT AGA CAT TC

IIPOL F GTT GAA AAC CGT TGT AG R GCT GAA CTT AAT CTC CT

12 F GAG AAT GTC TAA TGG TCC TC long12-15 F TCA GGC AGG AGA TTA AGT TCA

R GGT ACA TTA TGT GCG TGT AT

13A F ATA CAC GCA CAT AAT GTA CC

R TAT ATG GGC TGC TTG GTT GT

13B F AGA CAG GAT GTA GAC AGT GG

R AGT TGC ATG GAA CCA GAT

14 F TCA TTC AAA ATA TAC TGA CTG

R GAG AAG AGA AAT CAG AGA CAT

15 F AAA TCC CAG GTC CTT TAG C

R GTT ATT GTT TCT ACT GGC AAC T

8kb F AGT CTG TGC CAC AGG TTT GAA ATC GAA

F CAG GCT TGT ATG TCT GCG ACA ACG TTC

T A RE Fragmentsize(bp)

52 296

58 427

52 HindIII 206,294

60 Hinfl 197,303,92

56 390

52 301

52 AluI 326,145

56 267

52 311

56 444

52 329

58 EcoRl 463, 169

56 DdeI 178, 251, 150

52 HindllI 265, 385 52 Alul 44 ,98 ,285

48 85

56 Haelll 377,255 54 2334

59 291

59 350

52 490

56 Pstl 262,426

50 8181

TA, Annealing temperature in °C; R E , restriction enzyme applied; F, forward; R, reverse.

2.4. Data analysis

Allele frequencies were determined by direct gene counting. A maximum likelihood tree was constructed using the CONTML program from Phylip package 3.5 (Felsenstein, 1993). Heterozygosity hi at each sequence

a

site i was calculated as: hi= ~ p / 1 -p j ) , where pj is a

frequency of the jth allele and a is a number of alleles at the ith site. Microsatellite heterozygosity, H,, for the (T), site, was calculated accordingly and reported sepa- rately from H, calculated below. Average heterozygosity per nucleotide H, (nucleotide diversity; including here substitutions and small insertions/deletions) was calcu- lated as a sum of heterozygosities hi divided by the total

L

DNA length L(L=7622): H , = ~ hi~L; standard error: i = l

S E = [ i=l~(hi--H")2/(L--1)L] I/2"TheH"valuesf°rthe

groups of populations, such as Africans, non-Africans or the world, were obtained as above, using allele frequencies p~ averaged across the constituting subpopul- ations. Fsr statistics were calculated according to Hartl

Hr--H s and Clark (1989): F s r - - - , where Hr is the

Hr nucleotide diversity in the group of subpopulations and H s is computed by dividing the sum of nucleotide diversities of the constituting subpoputations by their number. Fsr values for individual polymorphic sites were obtained using h i rather than H, values to calculate Hr and Hs.

E Zictkiewicz et al. / Gene 205 (1997) 161-171

Table 3 Oligonucleotide probes used in ASO hybridization

165

Polymorphic A S O sequence T H Polymorphic site site

2A-3A F TCT GTG AAA CAG GTT 37 45

2C-3C F TCT GTG CCA CAG GTT

2C-3A-5 F TCT GTG CAA CAg/a GTT 48

8 F AGT AGC TAA a/cAG TGT 37 50

I0 F TTC t/cGA CTC TCA ATA 37 55

12 F Tga/g TGC AAA TGC ATG 38 58T-59

58A-59

14 F CTT c/tGG CTA CCT TCG 38

64A-65

15 F GAA ACA AGA GAt/g ATT 38 64C-65

18 F ATC CAt/c TGC TTC TTA 38 70

20 F ATG CTT GCT TGG a/gCA 43 71

25 F ATG Cc/gC CAG TTG ATT 38 72

30 F AGG TAA ACA TAc/t AGC 43 85

32 F ATA At/cT TCT CTG TGG 37

86

33 F AAT GTt gtG TGT ACA 37

F AAT GT- --G TGT ACA TGC 87

35 R AGT TTC c/tTG CAT TTG 37 88

38 F GTT GTC arc ATT ATA 37 90

F GTT GTC --- ATT ATA TTA

93

40 F ATA Ga/gG AAA CAG CAT 37

95

ASO sequence T H

R GGC c/aTT AAA AAT TGG 37

R TCT ATA TCT Ac/tC CCA 37

F AGG TTA Ac/tT AGG GAG 37

F CCA TGT AGT Ata/g TAT 37

R CCA Ac/aT TTT ACT TCT 37

R CCA Ac/aT TAT ACT TCT

F AAT ATT CTC AAc/a CCT 37

R AGG t/gCT GAG AAT ATT

F TTT AAC Ag/aA AGC CTC 37

F AAC CGt/c TGT AGC ATA 37

R AGA CAA a/tAT ATA TGC 37

F CTG GTC TAG ATC TC ........ C 37

F G ATC TCt aga tct cC

F GTG AAA AAA a/gTC AAC 37

F TGG AGA Ac/tC TCT CTA 37

F CAT TTT TTT ACC t/cGA 37

F ATC CAC CAA AAc/t AGT 38

F ACT GAA TTA Tc/tG TCT 38

F CTA Tg/a/cA TAG AAT GAG 35

TH, Hybridization temperature in °C; F and R, oligonucleotides complementary to the - and + strand, respectively; allelic positions within the oligonucleotide sequence are in bold; whenever possible allelic probes are shown as a single sequence within variants in lower case separated by '/ '.

3. Results and discussion

3.1. DNA sequence variants and their frequency profiles in populations

An 8 kb region spanning exon 44 in the dystrophin gene (dys44, Fig. 1) was examined for the presence of DNA polymorphisms by combined SSCP/heteroduplex analysis (Fig. 2(a)). On average 20 chromosomes from

each of the 13 subpopulations listed in Table 1 were investigated, which amounts to a survey of almost 2 Mb DNA for the presence of mutations. ASO hybridization (Fig. 2(b)) was used to determine allele frequencies in the extended, global population sample (Table 1). The characterization of gel mobility variants (Fig. 2(a)) by DNA sequencing revealed a nine-allele (T)14_23 length variation (Fig. 2(c)) and 35 'point mutation' polymor- phisms due to l l transversions, 22 transitions, two

166

A

N121 D232

E. Zigtkiewicz et al. / Gene 205 (1997) 161-171

1 1 2 2 2 3 2 3 1 1 1 2 2 3 1 3 3 1 2 3

B AGG TTA ACT AGG GAG AGG TTA ATT AGG GAG

C

1 2 3 4 5 6 1 2 3 4 5 6 A ~ Q ........ i A i B B : c D ...... D F ~ E FE c c

15 14151619171516 15 2015 22151420 2117 221522 15 1522151517

DS

Fig. 2. Polymorphism detection and genotyping. (A) SSCP/heteroduplex analysis (moblity variants are denoted at the top, ND is the non-denatured control, DS indicates migration of double-stranded DNA); (B) ASO dot-blot hybridization (positions B5 and G6 correspond to 'no DNA' controls; the sequences of ASO probes are shown at the top); (C) denaturing gel electrophoresis of (T)n variants (n is indicated at the top).

3-nucleotide deletions and one 8-nucleotide duplication (Table 3). The ratio of substitutions to small insertions/deletions was approximately 10, and that of transitions to transversions was 2.

Frequency data (excluding length polymorphism) were used to construct a maximum likelihood tree (Felsenstein, 1993): the African populations clustered together separately from the non-African ones (Fig. 3). The ancestral state of each polymorphic site was inferred by comparison with the orthologous non-human pri- mates: allele shared with at least two great apes, i.e. chimpanzee, gorilla and/or orangutan, was considered ancestral (Zietkiewicz et al., submitted). A theoretical ancestral population obtained by setting ancestral allele frequencies at 1.0 (see Batzer et al., 1994; Mountain and Cavalli-Sforza, 1994; Nei and Takezaki, 1996)joined the tree between M'Buti Pygmies and other African populations, consistent with the African origin of modern humans (Stringer and Andrews, 1988; Lahr and Foley, 1994; Tattersal, 1995).

The number of segregating sites (Table 1) was typi- cally higher in African subpopulations than in non- African ones and ranged from 15 in Karitiana, a small inbred Amerindian tribe (Kidd et al., 1991), to 31 in

AfricanAmer R i m a i b e ~

NigeriM~sY

M a y ~ Karitiana

European~ / ~ e

~ S i b e r i a n Japanese. 77r'-- ~ Chinese k ~ PapuanNG

M'Buti

Fig. 3. Maximum likelihood tree of human populations. The genetic distances are proportional to the branch lengths; bootstrap values (100 replicates) are shown at the branching points; the putative root (indi- cated by an arrow) was obtained by setting the ancestral allele frequen- cies at one (see text).

African Americans, a highly heterogeneous and admixed group (Chakraborty et al., 1992). The frequency profiles of polymorphisms in populations are compared in a

E. Zif tkiewicz et al. / Gene 205 (1997) 161 171 167

series of histograms in Fig. 4. Fifteen segregating sites, with an overall new (i.e. non-ancestral) allele frequency > 0.1 were observed in all the populations. We conclude that these polymorphisms had to be present in an ancestral population, which gave rise to both African and non-African groups. Thirteen polymorphisms were endemic to Africa and only four such sites were found outside Africa (new alleles at positions '10' and '85', frequent in Europeans, Siberians and Amerindians,

always occurred together; in African Americans they were considered as European admixture). These African and non-African polymorphisms were rare, with new allele frequencies < 0.1. Four out of the 13 African and two out of the four non-African polymorphisms were 'private ' , i.e. a new allele was present in a single popula- tion only (indicated by asterisks in Fig. 4). The polymor- phism at site '55' was found only in Europeans and African Americans. A new allele was fixed at site '30'

1.0

0.8

0.6

0.4

0.2

0.0

A,r,,m,,. ,...."ltlllll . . . . . . : . . . , . . . . . , . , . . . . . . , , I . . . t . . . . . . . . . . . . . . . .

-,,,,",,lllltlii • * i . . i l i ! i l l ,,,,,,,tlllllll ,,,,I .,h

. , . . . . . o . . . . . • . . . . . ° . . . . . , . . . . . , . . . . . , . . . . .

• . . . . . . . . . . , . , ' . . ' ' ' r ' . l . ' . . . . • . . . . . • . . . . .

I .,a~a tii . ..... _._._ ...,,,I. !.0.8 il !1 I/I ! il,ll

- ilrll : I1,,,,,,,, I r, . , . . . . . , . . . . - . . . . . . ° . . . . . ° . . . . . , . . . . . • . . . . .

European Illlllllllll . . . . . . . , . . . . . . . 0 a . . J i . ! , , . . . . . . . . . . . . . .

Siberian - j~ l

,.,..llllltlllllll . in.e '""'llll'llli

. . . . . . . . . o . . . . . . . . . . . . . . . • . . . . . • . . . . .

, o . . . . . . . . ° . . . . . . . . . . . . . . • . . . . . • . . . . .

Maya

.aritana llllli lil,,,, 1.0 001 0•0 • g i l I . . . . . . . • - . . . . . . . . . . . . • . . . . . • . . . . .

Fig. 4. Distribution of dys44 polymorphisms among populations. Histograms of new allele frequencies are ordered according to their increasing abundance in the world population. Globally distributed polymorphisms are represented by dark grey bars, endemic African *positions in black and those found only in non-Africans in light grey. Asterisks indicate polymorphisms found in a single population. Arbitrary names of the polymorphisms (Table 3) are shown at the bottom histograms only.

168 E. Ziftkiewicz et al. /Gene205 (1997) 161-171

in all non-African populations; at site '14' this occurred in some of the African populations, whereas at site '58' in a number of African and non-African populations (Fig. 4).

In Africans the profiles of new allele frequencies (Fig. 4) are similar in all the populations, except for the most distant M'Buti Pygmies (Fig. 3). In Biaka Pygmies we noted the shortage of rare new alleles; in Mossi, Rimaibe, Nigerians and Biaka Pygmies the new allele was fixed at position '58' whereas in Mossi, Nigerians and M'Buti Pygmies this occurred at position '14'. In non-Africans, Fig. 4 reveals differences among popula- tions at a number of positions. In Chinese, Japanese and Papua New Guineans the new alleles '95-1' and '14' are fixed whereas those at positions '10' and '85' are missing. Papua New Guineans share with the Amerindian populations a higher level of new alleles at positions '8', '2' and '71'. Maya and Karitiana also share new alleles at positions '10' and '85'; these are also present at significant frequency in Europeans and Siberians. Otherwise the frequency profiles in Amerindians appear different; the distinct one in Karitiana may be ascribed to the effect of the genetic drift expected to be more pronounced in a small and inbred population (Kidd et al., 1991). Due to their physical proximity dys44 polymorphisms are in linkage disequilibrium, which can be seen as correlated changes in certain allele frequencies within populations. However, based on tests by Tajima (1989) and Kimura and Crow (1964), these sites appeared as independent, neutral polymorphisms (data not shown).

Four to nine alleles were observed at the ( T ) n micro- satellite in African populations and from two to four in non-Africans (Fig. 5); a similar pattern to that seen in Fig. 4, but even more pronounced. The (T)~s allele had the highest frequency in all but one (M'Buti Pygmies) population; it could be considered the oldest or the most stable among the inter-converting length variants.

3.2. ~ r

Fsr values, reflecting the contribution of differences among subpopulations to the total divergence, strongly depend on how these subpopulations are defined within the total population (Fig. 6). The Fsr value of 0.147 characterized the world composed of 13 subpopulations. When the world population was considered to be com- posed of only two groups, Africans and non-Africans, the corresponding value was 0.036. This indicated that most of the divergence within the world was caused by differences among particular populations and not by the difference between Africans and non-Africans. Furthermore, the Fsr values for Africans and non- Africans were 0.072 and 0.158, respectively. The same overall tendencies were observed when Fsr was calcu- lated for each polymorphic site separately (Fig. 6). In

~ r h ~e~c !

24 19 14 21

17 22 16 20 15

allele (n)

1.0

0.8 = 0.6 D"

0.4 ~" o 0.2 _~ 0.0

• 5

i ; ' "

i =

ii / ° ,~.,a,~o'p ~ ~ / 0.6

/0.4 e. s~be~

0.2 "~ C~l~e se ,o.o . . . . . . . . .

pI~ 15 N I ~ ~ '~ 2 2 16 20 ~ ' t ~ - ' ...... ' a 14 21 17 lk[

24 "" allele (n)

Fig. 5. Allele frequencies of the (T), microsatellite among Africans and non-African populations (n, 14 to 24, are shown at the bottom).

non-Africans the Fsr values were much higher at the sites considered to represent old polymorphisms, whereas in Africans the Fsr values were similar across most of the polymorphic sites, except for the African- specific positions '12', '72' and '30'.

In Africans the major contribution to their overall divergence comes from M'Buti Pygmies: excluding this population resulted in lowering the African Fsr from 0.072 to 0.027. Among non-Africans the two groups contributing most to the Fsr value were Papua New Guineans and Karitiana. When both these populations were omitted from the calculation, the Fsr dropped from 0.158 to 0.059 (excluding only Papua New Guineans or Karitiana resulted in Fsr of 0.127 or 0.121, respectively). Nevertheless, in spite of these different representations of non-Africans and Africans, Fsr remained higher outside Africa. This is understandable in the light of the difference in geographical areas occupied by Africans and non-Africans.

E. Zigtkiewicz et al. / Gene 205 (1997) 161-171 169

0.5

0.4

0.3

0.2

0.1

0.0

0.0

0.1

0.2

0.3

0.4

0.5

World (13 populations) 0.147

,,,,,,,,,,,,, ] I, I,,,,,,,, I i, j I,, . ..,,,,, i...,, ...,.,i.,

World (Africans vs non-Africans)

0.036

Africans 0.072

,,,,, ,,, ,, I I,,,, ,,,,,,, I,

Non-Africans 0.158

Fig. 6. FsT values at each of the 35 polymorphic positions (bars) and based on the average heterozygosty per nucleotide (numbers). (a) World composed of the 13 populations studied; (b) world composed of two groups, Africans and non-Africans; (c) Africans (six populations); and (d) non-Africans (seven populations).

In conclusion, o u r fiST data are consistent with earlier reports on a variety of markers (Relethford, 1995; Hammer et al., 1997) and suggest a higher level of gene exchange within Africa than among non-African popula- tions. This in turn may explain the greater number of polymorphic sites seen in Africa. However, the same effect could also result from a larger population size on the Old Continent (Relethford and Harpending, 1995; Stoneking et al., submitted). On the other hand, higher nuclear DNA variability among Africans cannot be taken by itself as an indication of the African origin of modern humans. African and non-African populations share all old segregating sites and differ in rare, new polymorphisms (as measured by the frequency of the new, non-ancestral allele). These results are therefore consistent with a scenario where Africans and non- Africans diverged recently from a common ancestral population, with more opportunities (see above) to accumulate new variants existing subsequently in Africa than outside (Zietkiewicz et al., submitted).

3.3. Nucleotide diversity and the ascertainment bias

The density of segregating sites reported previously for the X chromosome ranged from 1/732 to 1/190 (Cooper et al., 1985; Hofker et al., 1986; Zietkiewicz et al., 1995). In dys44, with 35 sites within 7622 bp, this density is 1/218 and is probably an underestimate given the limitations of SSCP/heteroduplex analysis (Hayashi, 1992; Sheffield et al., 1993). Interestingly, using 34

restriction enzymes commonly applied in the earlier RFLP studies (e.g. Cooper and Schmidtke, 1984; Cooper et al., 1985; Antonarakis et al., 1988), five variant positions would be revealed within 540 bp of dys44 screened in that way. The resulting polymorphism den- sity of 1/108 would therefore be twice as large as our SSCP-based estimate. If only Europeans were analysed, the RFLP density would have been 1/180, compared to 1/382 (20/7622) based on the SSCP screening of the same population.

In our data set the H, values follow the same relative tendency as the number of polymorphic sites S (Table 1). H, is estimated at 0.101% for the world and its highest values are seen in Africans. When dys44 polymorphisms are ascertained in European DNAs only (Fig. 4), the resulting biased heterozygosity values H,~,r are lower. A significant effect of such a bias is only seen in African populations (Table 1); but even there, it does not exceed 25%. This is because European polymorphisms represent practically all common fre- quent sites, whereas rare new polymorphisms, particu- larly abundant in Africa, increase S but contribute little to the overall heterozygosity. However, if only European markers are considered, the ratio of nucleotide diversity in Africa and non-Africa becomes inverted (in Africans H,e.r of 0.084% is lower than in non-Africans, 0.09%). Furthermore, by typing only European markers we inadvertently lose the qualitative and quantitative infor- mation concerning polymorphic sites endemic to Africans or to other non-European populations (Fig. 4).

170 E. Zi~tkiewicz et al. /Gene205 (1997) 161 171

The knowledge of such polymorphisms is necessary to reconstruct the evolutionary history of our species. Those that can be considred 'private', may be important for detecting gene flow and reconstructing demographic history and relations among particular populations. Unbiased ascertainment of the polymorphisms thus appears to be essential to recognize the differences among the diversified human groups. In contrast to early nuclear data, mitochondrial DNA studies from the beginning included globally distributed human samples (Cann et al., 1987); this can be at least partly responsible for the conflicting portraits of human origins emerging from these two types of data (Hey, 1997). The microsatellites (Bowcock et al., 1994) and minisatellites (Armour et al., 1996) do not suffer from ascertaiment bias to the same extent (Rogers and Jorde, 1996), since these sites are often conserved across species (Zietkiewicz et al., 1994) so that their genomic distribu- tion as polymoprhic sites is usually pan-specific. The microsatellite heterozygosity Hm values in dys44 (Table 1) are higher in Africa; this is related to the greater number of (T)n alleles, and resembles the nucleo- tide diversity in a sense that finding a new microsatellite allele is like adding a new polymorphic site in the analysed sequence segment. However, we need nuclear markers with different characteristics, and ideally their combinations (Batzer et al., 1994, Tishkoff et al., 1996) to study the evolutionary origin of Homo sapiens, which thus requires samples from different and globally distrib- uted human populations (Cavalli-Sforza et al., 1991; Cavalli-Sforza et al., 1994).

4. Conclusions

(1) European bias leads to an inadvertent loss of the important genomic record concerning the recent evolution of our species. Therefore, it is essential to ascertain nuclear polymorphisms in a worldwide- distributed human population.

(2) Our data are consistent with the recent common origin of modern humans and indicate that differ- ences between Africans and non-Africans are a result of the recent history of these populations. African populations staying on a single continent remained in closer contact, whereas those outside, being dispersed over larger geographical areas, were more isolated, which reduced opportunities for exchange of the genetic material.

Acknowledgement

Wojciech Makalowski and Nur Akalin participated in primer design and early experiments, Guy Cardinal and Miguel Angel Roman in DNA sequencing. This

research was supported by a grant from the Canadian Genome Analysis and Technology Program (GO-12782).

References

Antonarakis, S.E., Oettgen, P., Chakravarti, A., Halloran, S.L., Hudson, R.R., Feisee, L., Karathanasis, S.K., 1988. DNA polymor- phism haplotypes of the human apolipoprotein APOA1- APOC3-APOA4 gene cluster. Hum. Genet. 80, 265-273.

Armour, J.A.L., Anttinen, T., May, C.A., Vega, E.E., Sajantila, A., Kidd, J.R., Kidd, K.K., Bertranpeti, J., Paabo, S., Jeffreys, A.J., 1996. Minisatellite diversity supports a recent African origin for modern humans. Nature Genet. 13, 154-160.

Batzer, M.A., Stoneking, A., Alegria-Hartman, M., Bazan, H., Kass, D.Y., Shaikh, T.H., Novick, G.E., Ioannou, P.A., Scheer, W.D., Herrera, R.J., Deininger, P.L., 1994. African origin of human specific polymorphic Alu insertions. Proc. Natl. Acad. Sci. USA 91, 12288-12292.

Bird, P.A., 1987. GpG islands as gene markers in the vertebrate nucleus. Trends Genet. 3, 342 347.

Bowcock, A.M., Kidd, J., Mountain, J.L., Hebert, J.M., Carotenuto, L., Kidd, K.K., Cavalli-Sforza, L.L., 1991. Drift, admixture, and selection in human evolution: a study with DNA polymorphisms. Proc. Natl. Acad. Sci. USA 88, 839-843.

Bowcock, A.M., Ruiz-Lineares, A., Tomfohrde, J., Minch, E., Kidd, J.R., Cavalli-Sforza, L.L., 1994. High resolution of human evolu- tionary trees with polymorphic microsatellites. Nature 368,455-457.

Cann, R.L., Stoneking, M., Wilson, A.C., 1987. Mitochondrial DNA and human evolution. Nature 325, 31-36.

Cavalli-Sforza, L.L., Wilson, A.C., Cantor, C.R., Cook-Deegan, R.M., King, M.-C., 1991. Call for a worldwide survey of human genetic diversity: a vanishing opportunity for the human genome project. Genomics 11,490 491.

Cavalli-Sforza, L.L., Menozzi, P. and Piazza, A. (1994) The history and geography of human genes. Princeton University Press, Princeton, New Jersey.

Chakravarti, A., Elbein, S.C., Permutt, M.A., 1986. Evidence for increased recombination near the human insulin gene: implication for disease association studies. Proc. Natl. Acad. Sci. USA 83, 1045-1049.

Chakraborty, R., Kamboh, M.I., Nwankwo, M., Ferrell, R.E., 1992. Caucasian genes in American Blacks: new data. Am. J. Hum. Genet. 50, 145-155.

Cooper, D.N., Schmidtke, J., 1984. DNA restriction fragment length polymorphisms and heterozygosity in the human genome. Hum. Genet. 66, 1-16.

Cooper, D.N., Smith, B.A., Cooke, H.J., Niemann, S., Schmidtke, J., 1985. An estimate of unique DNA sequence heterozygosity in the human genome. Hum. Genet. 69, 201-205.

Felsenstein, J. (1993) PHYLIP (Phylogeny Inference Package) version 3.5p. Distributed by the author. Department of Genetics, University of Washington, Seattle. [Felsenstein, J. (1989) PHYLIP-Phylogeny Inference Package (version 3.2), Cladistics, 5, 164-166.]

Fullerton, S.M., Harding, R.M., Boyce, A.J., Clegg, J.B., 1994. Molec- ular and population genetic analysis of allelic sequence diversity at the humanfl-globin locus. Proc. Natl. Acad. USA 91, 1805 1809.

Hammer, M.F., Spurdle, A.B., Karafet, T., Bonner, M.R., Wood, E.T., Novelletto, A., Malaspina, P., Mitchell, R.J., Horai, S., Jenkins, T., Zegura, S.L., 1997. The geographic distribution of human Y chro- mosome variation. Genetics 145, 787-805.

Hartl, D.L. and Clark, A.G. (1989) Principles of Population Genetics, Sinauer Associates, Sunderland, Massachusetts.

Hayashi, K., 1992. PCR-SSCP: A simple and sensitive method for

E. Zictkiewicz et al. / Gene 205 (1997) 161-171 171

detection of mutations in the genomic DNA. PCR Methods Appl. 1, 34-38.

Hey, J., 1997. Mitochondrial and nuclear genes present conflicting portraits of human origin. Mol. Biol. Evol. 14, 166-172.

Hofker, M.H., Skraastad, M.I., Bergen, A.A.B., Wapenaaar, M.C., Bakker, E., Millington-Ward, A., van Ommen, G.J.B., Pearson, P.L., 1986. The X chromosome shows less genetic variation at the restriction sites than the autosomes. Am. J. Hum. Genet. 39, 438-451.

Kidd, J.R., Black, F.L., Weiss, K.M., Balazs, I., Kidd, K.K., 1991. Studies of three Amerindian populations using nuclear DNA poly- orphisms. Hum. Biol. 63, 775-794.

Kimura, M., 1983. The neutral theory of molecular evolution. Cam- bridge University Press, Cambridge.

Kimura, M., Crow, J.F., 1964. The number of alleles that can be maintained in a finite population. Genetics 49, 725-738.

Labuda, D., Striker, G., 1989. Sequence conservation in Alu evolution. Nucleic Acids Res. 17, 2477-2491.

Lahr, M.M., Foley, R., 1994. Multiple dispersals and modern human origins. Evol. Anthropol. 3, 48-60.

Li, W.H., Sadler, L.A., 1991. Low nucleotide diversity in man. Genet- ics 129, 513-523.

Mountain, J.L., Cavalli-Sforza, L.L., 1994. Inference of human evolu- tion through cladistic analysis of nuclear DNA restriction polymor- phisms. Proc. Natl. Acad. Sci. USA 91, 6515-6519.

Murray, J.C., Mills, K.A., Demopulos, C.M., Hornung, S., Motulsky, A.G., 1984. Linkage disequilibrium and evolutionary relationships of DNA variants (restriction enzyme fragment length polymor- phisms) at the serum albumin locus. Proc. Natl Acad. Sci. USA 81, 3486-3490.

Nei, M., Takezaki, N., 1996. The root of the phylogenetic tree of human populations. Mol. Biol. Evol. 13, 170-177.

Orita, M., Iwahana, H., Kanazawa, H., Hayashi, K., Sekiya, T., 1989. Detection of polymorphisms of human DNA by gel electrophoresis

as single-strand conformational polymorphisms. Proc. Natl. Acad. Sci. USA 86, 2766-2770.

Relethford, J.H., 1995. Genetics and modern human origins. Evol. Anthrop. 4, 53-63.

Relethford, J.H., Harpending, H.C., 1995. Ancient differences in pop- ulation size can mimic a recent African origin of modern humans. Curr. Anthropol. 36, 667-675.

Rogers, A., Jorde, L.B., 1996. Ascertainment bias in estimates of average heterozygosity. Am. J. Hum. Genet. 58, 1033-1041.

Sheffield, V.C., Beck, J.S., Kwitek, A.E., Sandstrom, D.W., Stone, E.M., 1993. The sensitivity of single-strand conformation polymor- phism analysis for the detection of single base substitutions. Geno- mics 16, 325-332.

Stringer, C.B., Andrews, P., 1988. Genetic and fossil evidence for the origin of modern humans. Science 239, 1263-1268.

Tajima, F., 1989. Statistical method for testing the neutral mutation hypothesis by DNA polymorphism. Genetics 123, 585-595.

Tattersal, I. (1995) The Fossil Trail. Oxford University Press, New York, Oxford.

Tishkoff, S.A., Dietzsch, E., Speed, W., Pakstis, A.J., Kidd, J.R., Cheung, K., Bonnr-Tamir, B., Santachiara-Benerecetti, A.S., Moral, P., Krings, M., P~i/ibo, S., Watson, E., Risch, N., Jenkins, T., Kidd, K.K., 1996. Global patterns of linkage disequilibrium at the CD4 locus and modern human origins. Science 271, 1380 1387.

Zietkiewicz, E., Akalin, N., Labuda, D., 1995. Neutral polymorphisms in the deletion-prone regions of the dystrophin gene. Hum. Hered. 45, 80-88.

Zietkiewicz, E., Sinnett, D., Richer, C., Mitchell, G., Vanasse, M., Labuda, D., 1992. Single strand conformational polymoprhisms (SSCP): detection of useful polymorphisms at the dystrophin locus. Hum. Genet. 89, 453-456.

Zietkiewicz, E., Rafalski, A., Labuda, D., 1994. Genome fingerprinting by simple sequence repeat (SSR)-anchored polymerase chain reac- tion amplification. Genomics 20, 176-183.